http://reviews.llvm.org/D21564
This patch is an initial implementation for #distribute parallel for.
The main differences that affect other pragmas are:
The implementation of 'distribute parallel for' requires blocking of the associated loop, where blocks are "distributed" to different teams and iterations within each block are scheduled to parallel threads within each team. To implement blocking, sema creates two additional worksharing directive fields that are used to pass the team assigned block lower and upper bounds through the outlined function resulting from 'parallel'. In this way, scheduling for 'for' to threads can use those bounds.
As a consequence of blocking, the stride of 'distribute' is not 1 but it is equal to the blocking size. This is returned by the runtime and sema prepares a DistIncrExpr variable to hold that value.
As a consequence of blocking, the global upper bound (EnsureUpperBound) expression of the 'for' is not the original loop upper bound (e.g. in for(i = 0 ; i < N; i++) this is 'N') but it is the team-assigned block upper bound. Sema creates a new expression holding the calculation of the actual upper bound for 'for' as UB = min(UB, PrevUB), where UB is the loop upper bound, and PrevUB is the team-assigned block upper bound.
llvm-svn: 273705
-Wfor-loop-analysis warnings for a for-loop with a condition variable. In such
a case, the loop condition variable is modified on each iteration of the loop
by definition.
Original commit message:
Rearrange condition handling so that semantic checks on a condition variable
are performed before the other substatements of the construct are parsed,
rather than deferring them until the end. This allows better error recovery
from semantic errors in the condition, improves diagnostic order, and is a
prerequisite for C++17 constexpr if.
llvm-svn: 273600
During the core analysis, ExplodedNodes are added to the
ExplodedGraph, and those nodes are cached for deduplication purposes.
After core analysis, reports are generated. Here, trimmed copies of
the ExplodedGraph are made. Since the ExplodedGraph has already been
deduplicated, there is no need to deduplicate again.
This change makes it possible to add ExplodedNodes to an
ExplodedGraph without the overhead of deduplication. "Uncached" nodes
also cannot be iterated over, but none of the report generation code
attempts to iterate over all nodes. This change reduces the analysis
time of a large .C file from 3m43.941s to 3m40.256s (~1.6% speedup).
It should slightly reduce memory consumption. Gains should be roughly
proportional to the number (and path length) of static analysis
warnings.
This patch enables future work that should remove the need for an
InterExplodedGraphMap inverse map. I plan on using the (now unused)
ExplodedNode link to connect new nodes to the original nodes.
http://reviews.llvm.org/D21229
llvm-svn: 273572
The PIC and PIE levels are not independent. In fact, if PIE is defined
it is always the same as PIC.
This is clear in the driver where ParsePICArgs returns a PIC level and
a IsPIE boolean. Unfortunately that is currently lost and we pass two
redundant levels down the pipeline.
This patch keeps a bool and a PIC level all the way down to codegen.
llvm-svn: 273566
are performed before the other substatements of the construct are parsed,
rather than deferring them until the end. This allows better error recovery
from semantic errors in the condition, improves diagnostic order, and is a
prerequisite for C++17 constexpr if.
llvm-svn: 273548
Add support for /Ob1 (and equivalent -finline-hint-functions), which enable
inlining only for functions marked inline, either explicitly (via inline
keyword, for example), or implicitly (function definition in class body,
for example).
This works by enabling inlining pass, and adding noinline attribute to
every function not marked inline.
Patch by Rudy Pons <rudy.pons@ilod.org>!
Differential Revision: http://reviews.llvm.org/D20647
llvm-svn: 273440
This is a fix for PR28112:
https://llvm.org/bugs/show_bug.cgi?id=28112
The FP comparison intrinsics that take an immediate parameter (rather than specifying
a comparison predicate in the function name) were added with AVX; these are macros in
avxintrin.h. This patch makes clang behavior match gcc (error if a program tries to use
these without -mavx) and matches the Intel documentation, eg:
VCMPPS: m128 _mm_cmp_ps(m128 a, __m128 b, const int imm)
'V' means this is intended to only work with the AVX form of the instruction.
Differential Revision: http://reviews.llvm.org/D21306
llvm-svn: 273311
Summary:
Added calculateRangesAfterReplaments() to calculate original ranges as well as
newly affacted ranges in the new code.
Reviewers: klimek, djasper
Subscribers: cfe-commits, klimek
Differential Revision: http://reviews.llvm.org/D21547
llvm-svn: 273290
Include opencl-c.h by default as a module to utilize the automatic AST caching mechanism of clang modules.
Add an option -finclude-default-header to enable default header for OpenCL, which is off by default.
Differential Revision: http://reviews.llvm.org/D20444
llvm-svn: 273191
Add -mno-iamcu option to:
1) Countervail -miamcu option easily
2) Be compatible with GCC which supports this option
Differential Revision: http://reviews.llvm.org/D21469
llvm-svn: 273147
This mirrors the many other -i*after options to insert a new system search
directory at the end of the search path. This makes it possible to actually
inject a search path after the resource dir. This option is similar in spirit
to the /imsvc option in the clang-cl driver. This is needed to properly use the
driver for Windows targets where the clang headers wrap some of the system
headers.
This concept is actually useful on other targets (e.g. Linux) and would be
really easy to support on the core toolchain.
llvm-svn: 273016
Fix a regression which forbids using -std=cl|CL1.1|CL1.2|CL2.0 in driver.
Allow -std and -cl-std={cl|CL}{|1.1|1.2|2.0}.
Differential Revision: http://reviews.llvm.org/D20630
llvm-svn: 273015
Reapplying patch in r272777 which was reverted
because the llvm patch which added support
for generating the mcrr/mcrr2 instructions
from the intrinsic was causing an assertion
failure. This has now been fixed in llvm.
llvm-svn: 272983
This patch fixes a bug where we'd segfault (in some cases) if we saw a
variadic function with one or more pass_object_size arguments.
Differential Revision: http://reviews.llvm.org/D17462
llvm-svn: 272971
This is the second patch required to support compilation for Intel MCU target (e.g. Intel(R) Quark(TM) micro controller D 2000).
When IAMCU triple is used:
* Recognize and use IAMCU GCC toolchain
* Set up include paths
* Forbid C++
Differential Revision: http://reviews.llvm.org/D19274
llvm-svn: 272883
Parameters and non-static members of aggregates are still excluded,
and probably should remain that way.
Differential Revision: http://reviews.llvm.org/D19754
llvm-svn: 272859
When static variables are used in inline functions in header files anything that
uses that function ends up with a reference to the variable. Because
RecursiveASTVisitor uses the inline functions in LambdaCapture that use static
variables any AST plugin that uses RecursiveASTVisitor, such as the
PrintFunctionNames example, ends up with a reference to these variables. This is
bad on Windows when building with MSVC with LLVM_EXPORT_SYMBOLS_FOR_PLUGINS=ON
as variables used across a DLL boundary need to be explicitly dllimported in
the DLL using them.
This patch avoids that by adjusting LambdaCapture to be similar to before
r263921, with a capture of either 'this' or a VLA represented by a null Decl
pointer in DeclAndBits with an extra flag added to the bits to distinguish
between the two. This requires the use of an extra bit, and while Decl does
happen to be sufficiently aligned to allow this it's done in a way that means
PointerIntPair doesn't realise it and gives an assertion failure. Therefore I
also adjust Decl slightly to use LLVM_ALIGNAS to allow this.
Differential Revision: http://reviews.llvm.org/D20732
llvm-svn: 272788
Patch adds intrinsics for mrrc/mrrc2. The
intrinsics for mrrc/mrrc2 return a single
uint64_t to represent two 32 bit values.
The mcrr/mcrr2 intrinsic was changed to
accept a single uint64_t instead of two
32 bit values as the input for consistency.
Differential Revision: http://reviews.llvm.org/D21179
llvm-svn: 272777
classes.
MSVC actively uses unqualified lookup in dependent bases, lookup at the
instantiation point (non-dependent names may be resolved on things
declared later) etc. and all this stuff is the main cause of
incompatibility between clang and MSVC.
Clang tries to emulate MSVC behavior but it may fail in many cases.
clang could store lexed tokens for member functions definitions within
ClassTemplateDecl for later parsing during template instantiation.
It will allow resolving many possible issues with lookup in dependent
base classes and removing many already existing MSVC-specific
hacks/workarounds from the clang code.
llvm-svn: 272774
Summary:
This is similar to other loop pragmas like 'vectorize'. Currently it
only has state values: distribute(enable) and distribute(disable). When
one of these is specified the corresponding loop metadata is generated:
!{!"llvm.loop.distribute.enable", i1 true/false}
As a result, loop distribution will be attempted on the loop even if
Loop Distribution in not enabled globally. Analogously, with 'disable'
distribution can be turned off for an individual loop even when the pass
is otherwise enabled.
There are some slight differences compared to the existing loop pragmas.
1. There is no 'assume_safety' variant which makes its handling slightly
different from 'vectorize'/'interleave'.
2. Unlike the existing loop pragmas, it does not have a corresponding
numeric pragma like 'vectorize' -> 'vectorize_width'. So for the
consistency checks in CheckForIncompatibleAttributes we don't need to
check it against other pragmas. We just need to check for duplicates of
the same pragma.
Reviewers: rsmith, dexonsmith, aaron.ballman
Subscribers: bob.wilson, cfe-commits, hfinkel
Differential Revision: http://reviews.llvm.org/D19403
llvm-svn: 272656
Summary:
The validity of ABI/CPU pairs is no longer checked on the fly but is
instead checked after initialization. As a result, invalid CPU/ABI pairs
can be reported as being known but invalid instead of being unknown. For
example, we now emit:
error: ABI 'n32' is not supported on CPU 'mips32r2'
instead of:
error: unknown target ABI 'n64'
Reviewers: atanasyan
Subscribers: sdardis, cfe-commits
Differential Revision: http://reviews.llvm.org/D21023
llvm-svn: 272645
If definition of default function argument uses itself, clang crashed,
because corresponding function parameter is not associated with the default
argument yet. With this fix clang emits appropriate error message.
This change fixes PR28105.
Differential Revision: http://reviews.llvm.org/D21301
llvm-svn: 272623
David Majnemer